- H-FARM AI's Newsletter
- Posts
- Z.ai ships GLM-5.2 with 1M-token coding context
Z.ai ships GLM-5.2 with 1M-token coding context
PLUS: OpenAI preps GPT-Bidi for ChatGPT Voice & Android 17 adds native MCP for AI agents. DeepSeek raises $7.4B at $50B valuation, Cursor ships Origin, a Git built for agents.

1️⃣ Z.ai releases GLM-5.2, an open-weights coding model with a 1M-token context window that rivals GPT 5.5 and Claude Opus 4.8 on benchmarks. 2️⃣ OpenAI is building GPT-Bidi, a bidirectional voice system that lets ChatGPT listen and speak simultaneously with multiple intelligence modes. 3️⃣ Google's Android 17 introduces AppFunctions and MCP support, enabling on-device AI agents to discover and execute app tools autonomously. |
|
H-FARM EVENTS ON AI
ClawCamp 2026 — The Personal AI Agents Bootcamp
ClawCamp 2026 is a week-long residential bootcamp where you go from "never touched an AI agent" to running your own Personal AI agent. It's an H-FARM College Summer Camp built with H-FARM AI and the Cisco Networking Academy, hosted on the H-FARM Campus in Roncade (Venice).
Save the date: 5–12 July 2026. Residential, in English, for curious minds 16–25+ — no advanced coding required. Just 15 seats per cohort, selected on a rolling basis.
Five hands-on build days on OpenClaw, the open-source framework for Personal AI Agents, plus a daily ClawTalk, a day in Venice with the cohort, and a final pitch on stage.
Every admitted student leaves with a working agent, a certificate, and a lifetime invite to the global ClawCon community.

MAIN AI UPDATES / 17th June 2026
🧑💻 Z.ai ships GLM-5.2 with 1M-token coding context 🧑💻
An open-weights coding model narrows the access gap with closed frontier systems.
Z.ai has released GLM-5.2, a powerful open-weights coding model featuring a 1 million-token context window and new reasoning controls including High and Max effort modes. The model supports long-horizon coding tasks across entire codebases and is competitive with GPT 5.5 and Claude Opus 4.8 on coding benchmarks — the closest an open model has come to matching top closed systems. Released under an MIT license, GLM-5.2 is immediately available to Coding Plan users, with API access, chatbot support, and full open weights rolling out next week. This directly pressures closed-model pricing across the coding tool market.
🎙️ OpenAI preps GPT-Bidi for ChatGPT Voice 🎙️
A bidirectional voice system enables real-time interruptions and multi-speed reasoning modes.
OpenAI is developing a major upgrade for ChatGPT Voice built around a new system called GPT-Bidi, which enables true bidirectional conversation. Unlike current voice assistants, the model can listen and speak simultaneously, handling interruptions and adjusting its response mid-sentence for more natural dialogue. References to GPT-Bidi have already surfaced across OpenAI's apps. The update also introduces High, Medium, and Instant intelligence modes, giving users control over the trade-off between deeper reasoning and faster responses. This could intensify competitive pressure in the voice AI space as rivals scramble to match real-time conversational fluidity.
📱 Android 17 adds native MCP for AI agents 📱
Google's latest Android rollout integrates MCP protocol for autonomous on-device agent workflows.
Google's Android 17 introduces AppFunctions and Android MCP (Model Context Protocol), enabling apps to expose orchestratable tools that on-device AI agents can discover and execute autonomously. Google positioned this as part of a broader shift toward an "intelligence system" with deep AI integration across the entire platform. This represents a major architectural evolution — moving Android from a traditional app-based OS to one designed natively for AI agent workflows, a shift that reshapes app distribution for developers. The integration could fundamentally change how users interact with their devices through AI intermediaries rather than manual app navigation.
INTERESTING TO KNOW
💰 DeepSeek raises $7.4B at $50B valuation 💰
DeepSeek raised more than $7.4 billion in a funding round valuing the company at over $50 billion, making it China's most valuable AI startup. Founder Liang Wenfeng personally invested around $3 billion, while a government-backed fund contributed approximately $150 million. The capital will fuel R&D and computing infrastructure expansion — a clear competitive pressure signal as pricing and capability battles heat up globally and DeepSeek's open-source models continue gaining international traction.
⚡ Cursor ships Origin, a Git built for agents ⚡
Cursor launched Origin, a new GitHub competitor designed as a replacement for traditional version control in agent-scale software development. The Git-compatible forge assumes numerous AI agents will be cloning, branching, committing, and reviewing code in parallel — a fundamental rethinking of how development infrastructure works. Alongside Origin, Cursor also released an iOS mobile app in beta, positioning itself as a full development platform directly challenging GitHub's dominance with an autonomous AI-driven workflow vision.

📩 Have questions or feedback? Just reply to this email , we’d love to hear from you!
🔗 Stay connected:
